infrastructure for connectivity and innovation: the conceptual … · 2015-11-20 ·...
TRANSCRIPT
Chapter 2
Infrastructure for Connectivity and Innovation:
The Conceptual Framework
ERIA CADP research team
November 2015
This chapter should be cited as
ERIA CADP research team (2015), ‘Infrastructure for Connectivity and Innovation: The
Conceptual Framework’, in ERIA (eds.), The Comprehensive Asia Development Plan 2.0
(CADP 2.0). ERIA Research Project Report 2014-04, Jakarta: ERIA, pp.7-30.
7
Chapter 2
Infrastructure for Connectivity and Innovation: The Conceptual Framework
This chapter discusses the theoretical background and the conceptual framework
for CADP 2.0. The original CADP (ERIA, 2010) placed the fragmentation theory and new
economic geography at the centre of the analytical approach. CADP 2.0 follows the same
path and at the same time further extends it to reflect recent changes in the development
stages of ASEAN and East Asia as well as the advancement of economic research at ERIA.
CADP 2.0 proposes the direction of infrastructure development not only for connectivity
but also for innovation.
The chapter plan is as follows: the first and second sections review and expand the
framework of the fragmentation theory and new economic geography. The third section
discusses innovation in industrial agglomeration as the microeconomic source of
productivity growth. The fourth section argues the implication of our development
strategy for the narrowing of geographical and industrial development gaps. The fifth
section links the conceptual framework to infrastructure development for connectivity
and innovation.
2-1. The Fragmentation Theory
2-1-1. Fragmentation and the second unbundling
Since the mid-1980s, the world economy has started using a new type of
international division of labour in production processes or tasks, instead of depending on
the traditional industry-by-industry division of labour. The fragmentation theory (Jones
and Kierzkowski, 1990) and the second unbundling (Baldwin, 2011) provide a conceptual
framework to understand the mechanics.
Figure 2.1 illustrates the fragmentation theory. Suppose that before the
fragmentation of production, a large factory took care of all production activities from
upstream to downstream. It was a factory, for example, in the electronics industry, which
was capital-intensive or human-capital-intensive so that it was located in a developed
country, following the traditional comparative advantage theory. If we carefully looked at
8
the factory, however, it included diversified production processes that used different
inputs and different technologies. Therefore, if we can separate some of the production
processes into production blocks and place them in appropriate locations, we may save
on total production cost. This is so-called fragmentation of production.
Figure 2.1. The Fragmentation Theory
Source: ERIA CADP research team.
Whether such fragmentation of production works depends on two conditions.
First, the savings in production costs in a fragmented production block should be large
enough. Second, costs of the service link that connects remotely located production
blocks must not be too high. Fragmentation is a powerful tool to exploit differences in
location advantages, particularly between countries/regions at different development
stages. It can be much more flexible and articulate than the traditional industry-wise
division of labour in taking advantage of gaps in factor prices, resource availability,
logistics arrangements, policy environments, and others. On the other hand, it must at
least partially overcome geographical distance by reducing service link costs, which
include transport costs in terms of monetary and time dimension, telecommunication
costs, and various coordination costs between production blocks.
9
The concept of the second unbundling further examines differences between
fragmented production and traditional industry-by-industry division of labour. The first
unbundling is the separation of production and consumption across national borders. It
started at the end of the 19th century with the introduction of the mass transport system,
such as steam ships and railways, and became a landmark for the formation of the world
economy dominated by the industry-by-industry international division of labour based on
comparative advantage. On the other hand, the second unbundling refers to the
international division of labour in terms of production processes and tasks. It was initiated
in the 1980s when the information and communications technology revolution drastically
reduced coordination costs in distance. Fundamental differences between the first and
the second unbundling reside in the way of dividing jobs/tasks with tight coordination
rather than differences across industries or between finished products and parts and
components. In the second unbundling, we have two-way flows of goods, ideas,
technology, capital, and technicians between remotely placed production blocks. This
requires a ‘trade–investment–services nexus’ supported by physical and institutional
connectivity. For the second unbundling, connectivity by logistics infrastructure must be
at a higher technical grade than for the first unbundling, which should take care of not
only monetary transport costs but also time costs and the reliability of logistic links.
As we will review in detail in Chapter 4, ASEAN and East Asia have been
forerunners in aggressively utilising the new international division of labour in their
development strategies. In particular, machinery industries are major players in extending
production networks. Machines consist of a large number of parts and components that
are produced by using diversified materials and technologies. The industry thus has a
sophisticated division of labour by nature and can be a natural forerunner of taking
advantage of production networks. We, of course, observe the development of
production networks or the second unbundling in other industries such as garment, food
processing, cut flowers, software, and others. However, most of these industries are still
in the traditional industry-by-industry international division of labour or the first
unbundling. The concept of global value chains has recently been popular (Elms and Low,
2013) but we have to be careful that the concept includes both the first and the second
unbundling. Production networks and the second unbundling are characterised by fast,
high-frequency, and synchronised transactions rather than slow, low-frequency, and less
10
coordinated transactions in a simplistic international input–output structure. The
development of production networks in machinery industries is actually a good indicator
for assessing the degree of participation in production networks by each country; that is,
because once the economic and policy environment allows machinery industries to
extend production networks, other industries can also do so.
2-1-2. Policies to reduce three kinds of costs
For a developing country to participate in production networks, it needs to find a
bottleneck. To join production networks, three kinds of costs need to be reduced: (i)
network set-up costs, (ii) service link costs, and (iii) production costs per se in production
blocks. If a country or a region has difficulty in joining production networks, some of these
costs are likely to be too high. That is the bottleneck. Then policymakers would like to
resolve the bottleneck by implementing necessary policies.
The government can reduce network set-up costs by policies to reduce investment
costs, such as the enhancement of stability, transparency, and predictability of
investment-related policies as well as the improvement of investment
facilitation/promotion services provided by foreign direct investment–hosted agencies
and industrial estates. Reduced service link costs may be achieved by a series of hard and
soft connectivity policies to overcome geographical distance and border effects, which
include the construction/operation of logistics infrastructure and trade
liberalisation/facilitation. Reduced production costs per se are realised by policies that
strengthen location advantages, which include, among others, enhancing and stabilising
supplies of economic infrastructure services for electricity and other utility supplies, as
well as industrial estate services.
2-1-3. Fragmentation and agglomeration
As a country or a region successfully participates in production networks and
accumulates a number of production blocks, industrial agglomerations will start to form.
Production networks in the world other than those in ASEAN and East Asia have barely
reached the stage of forming industrial agglomerations, and thus the parallel
development of fragmentation and agglomeration is not yet well recognised in the
11
academic literature. However, this is important, and the seemingly paradoxical
phenomenon can lucidly be explained by the extension of the fragmentation theory.
Kimura and Ando (2005) expand the fragmentation theory to two dimensions:
fragmentation in the geographical distance and fragmentation in the disintegration. The
new dimension, disintegration, means that fragmentation of production may occur in the
context of intra-firm or arm’s-length (inter-firm) division of labour. Arm’s-length division
of labour takes various forms of vertical linkages and outsourcing between unrelated firms.
Compared with intra-firm fragmentation, arm’s-length fragmentation is sensitive to
geographical distance. In particular, one side of a transaction is a local firm or a small or
medium enterprise in developing countries; the transaction is almost always in
geographical proximity to save on distance-sensitive transaction costs.
This is a dominant economic logic in forming agglomerations in ASEAN and East
Asia, which is quite different from typical cases in developed countries where industries
with high transport costs are attracted to the most immobile element, people.
Figure 2.2 illustrates the evolution of production networks. Production networks
typically start with a simplistic prototype as illustrated in Figure 2.2(a). This is just the
intra-firm fragmentation of production across national borders with back-and-forth
transactions between the United States (US) and Mexico, which is called cross-border
production sharing. Similar forms of production networks were observed in many places
at the beginning of the second unbundling era; examples are the semiconductor assembly
in Penang, Malaysia and garment operations between Hong Kong and Guangdong. As
production networks are extended and become sophisticated, fragmentation and
agglomeration start evolving at the same time. Figure 2.2(b) is the case of hard disk drive
production where geographical fragmentation dominates while industrial agglomeration
with arm’s-length transactions is initiated. Figure 2.2(c), on the other hand, is the case of
the automobile industry where the logic of industrial agglomeration dominates though
the formation is supported by parts and components supplies from abroad through
production networks.
12
Figure 2.2. The Evolution of Production Networks: Illustrations
Source: Ando and Kimura (2010), modified.
A firm in production networks actually combines four layers of transactions (Table
2.1). Layer 1 is a transaction within an industrial agglomeration where a just-in-time
system is literally operated. Layer 2 is a transaction within a subregion such as ASEAN that
is connected with middle-distance transportation, still sensitive to time costs. Layers 3
and 4 are transactions on a regional basis, such as in East Asia, and on a global basis, which
cannot be very time-sensitive anymore in most cases. The choice of four layers typically
depends on the elements presented in Table 2.2. Weights of four layers depend on the
economic and policy environment as well as industrial characteristics and corporate
strategies. In the case of the electronics industry, more weights are placed in long-
distance transactions because service link costs are low and arm’s-length transactions go
with relatively high credibility, balanced power, and are modular. On the other hand, the
automobile industry typically prefers short-distance transactions, particularly under a
corporate strategy like Toyota’s, because service link costs are typically high and arm’s-
length transactions go with relatively low credibility, unbalanced power, and total
integration.2
2 This view seems to be particularly applicable in the case of Toyota. On the other hand, some other automobile assemblers such as Volkswagen and Hyundai may apply more module interface as well as communised parts and components worldwide so that a system close to complete knockdown may apply. This issue has been investigated in the series of automobile industry studies by ERIA and Research Institute Auto Parts Industries, Waseda University (2014).
Secondary suppliers
Viet Nam
Headquarters or affiliate
'network''network'
13
Table 2.1. Four Layers of Transactions in Production Networks
Source: Originally in Kimura (2009), modified.
Table 2.2. Determinants of the Transaction Layer Choice
Source: Originally in Kimura (2009), modified.
The formation of industrial agglomerations calls for a new set of hard and soft
infrastructure—hard infrastructure for industrial agglomeration, and soft infrastructure
for reducing transaction costs in arm’s-length transactions.
2-2. New Economic Geography
2-2-1. Agglomeration and dispersion forces
New economic geography (Fujita, Krugman, and Venables, 1999; Baldwin, Forslid,
Martin, Ottaviano, and Robert-Nicoud, 2003) is another pillar of our conceptual
framework. It complements the fragmentation theory, particularly in considering ways of
participating in production networks. While the fragmentation theory inclines toward
individual firms’ decision-making, new economic geography looks at agglomeration forces
and dispersion forces generated by production–consumption interactions in both internal
Layer 1 Layer 2 Layer 3 Layer 4
(Within ind. agg.) (Within sub-region) (Within region) (Global)
Lead time Within 2.5 hours 1 to 7 days 1 to 2 weeks 2 weeks to 2 months
Typical transaction
frequencyMore than once in a day More than once a week One a week Less than once a week
Major transport mode Track Track/ship/airplane Ship/airplane Ship/airplane
Trip length Within 100km 100-1,500km 1,500-6,000km More than 6,000km
Layer 1 Layer 2 Layer 3 Layer 4
<Fragmentation (geographical)>
Network set-up costs (e.g., cost to invite upstream firms)
Service link costs (e.g., transport costs)
Location advantages (e.g., wages, economics of scale)
<Fragmentation (disintegration)>
Intra-firm vs. arm’s length (inter-firm)
In cases of intra-firm transactions:
Trust
Power balance
Architecture of firm-to-firm interface
Modular vs. total integration
small large
small
small
large
large
Intra-firmArm’s length (inter-firm)
weak strong
unbalanced balanced
integration modular
14
and external economies. In addition, new economic geography can think of a situation
where not only economic activities but also people (or labour) can move.
Figure 2.3 depicts the essence of new economic geography. Suppose we have a
core and a periphery in geographical distance. If trade costs between the core and the
periphery go down, both agglomeration forces and dispersion forces are generated.
Agglomeration forces mean that economic activities, people, and others are attracted to
the core where positive agglomeration effects are found in the form of the easiness of
finding business partners, the proximity to the market, and others. Positive agglomeration
effects are often formalised as a sort of economies of scale external to individual firms
that work within a certain geographical boundary. However, economies of scale internal
to individual firms may also work as a benefit from moving to the core. On the other hand,
dispersion forces generate movements of economic activities, people, and others from
the core to the periphery. One source of dispersion forces is negative agglomeration
effects or ‘congestion’ in the core, which includes wage increases, land price hikes, traffic
congestion, environmental pollution, and others. Some economic activities or people do
not like such congestion and may move from the core to the periphery. Another source
of dispersion forces is a difference in location advantages such as differences in wages
and others though this could also be interpreted as an element generated by ‘congestion’.
In contrast to a typical setting in Western Europe or the US where factor prices and other
location advantages do not differ much, the core and the periphery in ASEAN and East
Asia tend to have a large gap in development stages, factor prices, and others. We can
thus expect dispersion forces of considerable magnitude in our region in contrast to
situations in developed economies where agglomeration forces are almost always
dominant.
The fragmentation theory may naively recommend a reduction in service link costs
in order to participate in production networks. On the other hand, new economic
geography poses a caveat that a reduction in trade costs may generate both
agglomeration forces and dispersion forces; thus, we should properly control the two
forces to achieve a balanced growth between the core and the periphery.
15
Figure 2.3. New Economic Geography: Agglomeration Forces and Dispersion Forces
Source: ERIA CADP research team.
2-2-2. Supplementary policy package to control the two forces
How can we control the magnitude of agglomeration forces and dispersion forces?
In particular, when latecomers would like to join production networks, a certain
magnitude of dispersion forces must be generated, together with a reduction in trade
costs. The answer is to properly plan and implement supplementary policy package
together.
While the fragmentation theory also claims the necessity of enhancing location
advantages, the strength of new economic geography is its ability to consider both
agglomeration forces and dispersion forces as well as possible mobility of multiple
elements. Let us use the Mekong–India Economic Corridor (MIEC) for thought
experiments (Figure 2.4). MIEC is an economic corridor that connects Ho Chi Minh City,
Phnom Penh, Bangkok Metropolitan Area, and Dawei. It has great potential for being a
major manufacturing corridor in the near future. Think of the case of industrial
development in Phnom Penh and in Dawei.
16
Figure 2.4. Mekong–India Economic Corridor
Source: ERIA CADP research team.
The recent development of the so-called ‘Thailand+1’ investment is a good sign of
expanding production networks from the Bangkok Metropolitan Area to the neighbouring
countries, together with reducing service link costs or trade costs. However, things may
be a bit more complicated. As Figure 2.5 illustrates, the Bangkok Metropolitan Area has
recently attracted a substantial number of labour from neighbouring countries. In the case
of Cambodia, about 1 million out of 15 million Cambodians are now in Thailand working
in unskilled labour–intensive sectors and the informal sector rather than in Phnom Penh.
How can Phnom Penh attract labour from the rural areas and, at the same time, invite
production blocks from Thailand? This is the case where reduced trade costs make both
economic activities and people easier to move within Cambodia and across the national
border.
LAO PDR
VIET NAM
17
Figure 2.5. Scenario for the Development of Phnom Penh
Source: ERIA CADP research team.
Although we need to conduct a serious micro study on the profile of migrant
labour and its impact, a simplistic thought experiment is also useful. If the wage gap
between Bangkok and Phnom Penh is too large, people do not come to Phnom Penh
though production blocks may be motivated to come. On the other hand, if the wage gap
is too small, production blocks do not come though people may flow into Phnom Penh.
How can Phnom Penh attract both production blocks and people? The answer is the
improvement of location advantages and liveability in Phnom Penh. The supplementary
policy package may include the better provision of economic infrastructure services in
Phnom Penh including better special economic zones (SEZs), more stable supply of
electricity, and others. At the same time, people’s movement costs from the rural areas
to Phnom Penh may be reduced. People coming to Phnom Penh should be willing to stay
in Phnom Penh, even if the salary is a bit lower than in Thailand, and enjoy comfortable
living.
Another case of thought experiment is the Dawei development. Dawei also
intends to attract both production blocks and people, thus requiring more drastic
measures to meet its ambition than in the case of Phnom Penh. Currently, there is nothing
in Dawei but a vast industrial site. One of the challenges is how to attract labour. To
support this big industrial estate, we need at least half a million people. If some activities
Labour
Labour? Labour?
18
are labour-intensive, wages should still be lower than in the Bangkok Metropolitan Area.
This means that urban development just next to the industrial site is essential.
Figure 2.6. Scenario for the Development of Dawei
Source: ERIA CADP research team.
In addition, at least three projects—the industrial estate, highway connection to
Thailand, and a deep sea port—must be implemented at the same time (Figure 2.7). This
follows an old, yet important, idea of coordinated ‘big push’ (Rosenstein-Rodan, 1943;
Murphy, Shleifer, and Vishny, 1989). If we miss one of them, the feasibility of the whole
project would collapse.
On the other hand, as shown in Figure 2.8, Dawei may ‘leapfrog’ (Bresiz, Krugman,
and Tsiddon, 1993). Dawei is located far from the centre as a big vacuum. Land
reclamation and other project preparations may be easier than in mainland Myanmar. It
is closer to a massive industrial agglomeration in the Bangkok Metropolitan Area. It can
jump to modern industrial technology and just-in-time logistics links, rather than step-by-
step industrialisation. In this sense, speed will matter for Dawei. The construction of a
deep sea port would take at least 10 years though Thilawa and others might take more
time to have full industrialisation and a deep sea port. If so, Dawei could become a hub of
industrial activities and logistics, which would also play a leading role for the
industrialisation of mainland Myanmar.
labour
labour
labour
19
Figure 2.7. Economics of Coordinated Investments
Source: ERIA CADP research team.
Figure 2.8. The Theory of Leapfrogging
Source: ERIA CADP research team.
The Daweidevelopment
Deep sea port
Industrial estate
Highway connection
20
2-3. Industrial Agglomeration and Innovation
2-3-1. Catching-up and sources of technological information
In the globalisation era, local firms in developing countries inevitably face more
exposure to competition and, at the same time, may enjoy better access to advanced
technology. Up to some stages of development, the backwardness can potentially be an
advantage for developing countries to learn from advanced countries at relatively low
costs. Once we enter the era of the second unbundling, developing countries face
globalisation in a deeper way.
ERIA has continuously conducted micro-level studies with structured
questionnaires to scrutinise flows of technological information: what sort of technological
information is flowing from where to where. A local firm may have three channels to
access technology (Figure 2.9). The first is via affiliates of foreign firms in the same
industrial agglomeration that are often in the downstream of production networks. The
second is from universities or research institutes in the country. The third is direct learning
from abroad by exchanges of technicians or through exports and imports. According to
our questionnaire surveys, the first channel, via foreign affiliates in industrial
agglomeration, is dominant in ASEAN. Furthermore, a local firm that receives technical
training is likely to provide technical training to upstream firms (Kimura, Machikita, and
Ueki, 2015).
Figure 2.9. Three Channels to Get Access to Technology
Source: ERIA CADP research team.
21
This is quite different from old models of technology acquisition. In the cases of
Japan, South Korea, and Taiwan from the 1950s to 1970s, universities and research
institutes played substantive roles in technology transfers and spillovers. Learning for
export as well as imports of machines that embodied technology was also significant. In
the case of ASEAN, these channels are relatively weak, and links with foreign affiliates are
important. This indicates the weakness of indigenous capability of acquiring technology
and the possible benefits of fragmented production.
Of course, not all local firms are automatically eligible to participate in production
networks run by multinationals. To meet the strict quality standard of goods and services
requested by other firms in the higher tiers of a production network, local firms must clear
internal constraints—such as the lack of financial and managerial capability, weak
competitiveness, and difficulty in having wider information/networks—in addition to
external constraints, such as poor access to finance and unfavourable business and
investment climate (Vo, Narjoko, and Oum, 2010).
The development of small and medium enterprises (SMEs) is certainly important,
but we have to be careful as there exist different kinds of SMEs. In particular, SMEs in
cottage industries and those in supporting industries are quite different. The confusion of
these two may end up with inconsistent policies mixing social policy and economic policy.
Both types of SMEs should be promoted but for different purposes and with different
policy packages.
In connection with infrastructure development, it is crucial to form a critical mass
of industrial agglomeration to enhance opportunities for local firms to link with foreign
affiliates. Urban and suburban development is expensive and politically challenging but is
an essential part of economic development.
2-3-2. Process and product innovation
Productivity growth is derived from innovation at the micro level. In particular,
after reaching the middle-income level, the innovation capability of local firms becomes
the key for sustainable economic growth.
There is a ladder in innovation. The one at the lower end is process innovation. It
includes minor changes in production processes through kaizen and QC circles, the
improvement of production lines, and the restructuring of the whole operation. A firm
22
can improve efficiency while producing basically the same products or services. The higher
end of the ladder is product innovation. A firm here introduces a new product or service;
it could be new to the world, new to the country, new to the industry, or just new to the
firm.
The higher a firm moves up the innovation ladder, the greater internal capability
is required of it. The external interface of a firm also changes. The first channel of
technology acquisition, which is via affiliates of foreign firms, can work well for process
innovation and some product innovation for new to the firm. However, eventually, the
second and third channels are going to be important, particularly after reaching the upper
middle–income stage for product innovation. Then the supply of human resources will
become crucial (Figure 2.10).
Figure 2.10. Technology Acquisition and Product Innovation
Source: ERIA CADP research team.
23
2-3-3. Industrial agglomeration and urban amenities
Infrastructure development is crucial in upgrading innovation in ASEAN and East
Asia. The reason is twofold.
First, to accelerate process innovation and initiate some product innovation, the
formation of industrial agglomeration beyond a critical mass is essential. In the era of the
second unbundling, technology can come with fragmented production blocks, and local
firms should take advantage of the proximity. Once industrial agglomeration grows up to
a certain size, local firms have good chances to participate in production networks and get
access to technology (Figure 2.11).
Figure 2.11. SMEs and Industrial Agglomeration
SMEs = small and medium enterprises. Source: ERIA CADP research team.
In our experience in ASEAN and East Asia, a well-functioning industrial
agglomeration seems to be as large as a circle of 100-kilometre (km) diameter in the case
of a full-sized one and 50 km diameter in the case of a middle-sized one. The Bangkok
Metropolitan Area is one good example of a full-sized industrial agglomeration. The role
of government in the formation of an industrial agglomeration of the proper size is critical.
An industrial agglomeration must be supported by urban/suburban infrastructure, which
Managerial know-how
Access to technologyAccess to financeFostering human resourcesEstablishing industrial organisations
24
includes logistics connection with neighbouring industrial agglomerations through a large
port and an airport, urban and suburban highway system and urban transport, mass
supplies of economic infrastructure services such industrial estates, supplies of electricity,
water, and others.
By developing infrastructure, we must enhance positive agglomeration effects,
which allow firms to set up a just-in-time system within 2.5 hours and increase chances
for local firms to have a business relationship with multinationals. At the same time,
negative agglomeration effects should be reduced by slowing down wage hikes, keeping
living cost low, mitigating land speculation, avoiding traffic congestion, and staying away
from pollution problems.
Because huge positive and negative externalities result from agglomeration
effects, infrastructure projects for industrial agglomerations are often not financially
viable if they are implemented purely by the private sector. However, we still need to
implement some of them, with the involvement of the central/local governments that
provide partial subsidies or insurance.
Second, at the stage of active product innovation, we must nurture and attract
high-quality human resources and set up an innovation hub. It has not been much
discussed in ASEAN yet, but eventually we need to think of how to provide good urban
amenities or quality of life to attract intellectuals.
There are four critical urban amenities by Glaeser, Kolko, and Saiz (2001): (i) the
presence of a rich variety of services and consumer goods, (ii) aesthetics and physical
setting, (iii) good public services, and (iv) speed. The first urban amenity (i) must cover
something that even advanced Internet shopping cannot provide, (ii) includes intellectual
stimulus and comfortable living, (iii) contains opportunities for higher education and
safety, and (iv) means easiness to get around and acceptable length of commuting.
In most ASEAN Member States and other East Asian developing countries,
research and development (R&D) activities are still minimal. After the stage of upper-
middle income, it is important to strengthen universities and government research
institutes to accumulate R&D stock (Sunami and Intarakumnerd, 2011). Urban amenities
are essential to attracting intellectual people, and infrastructure development must be
headed in this direction.
25
2-4. The Narrowing of Development Gaps
Since Piketty’s Capital in the Twenty-first Century (2013) became a bestseller, the
issue of income distribution has been extensively discussed worldwide. Piketty claims that
the income share of the highest one percent population substantially increased in the past
few decades. Actually, such a pattern is observed in some countries. We at least cannot
immediately conclude that globalisation aggravates income disparity.
If we simply look at the Gini coefficients of income size distribution, China and
India have clearly experienced an upward trend since the 1990s whereas those of the
ASEAN Member States have recently increased or decreased, depending on the country.
Compared with that of other parts of the world such as Latin America, income disparity in
ASEAN is not very serious, with relatively high Gini coefficients in Malaysia and the
Philippines. In addition, the population below the poverty line has steadily reduced in
ASEAN.
How to deal with super-rich people will become an important political agenda in
ASEAN and East Asian countries at some point. In this aspect, Piketty’s claim of the
necessity of income redistribution policy would be applicable in the future. What ASEAN
should immediately confront, however, is income disparity due to development gaps that
are not pointed out by Piketty.
Development gaps are of two kinds: geographical and industrial.3 Geographical
development gaps are differences in income levels and development stages among
countries or among regions within a country. Industrial development gaps refer to
differences in productivity and development stages between multinationals and local
firms, between large firms and SMEs, or between manufacturing and non-manufacturing.
Our development strategy, if it works effectively, can narrow these two
development gaps. Geographical development gaps can be reduced in two ways. The first
is through fragmentation of production. This is particularly applicable when production
blocks move from a higher income country to a lower income country. If a less developed
country/region can attract production blocks and participate in production networks,
geographical development gaps are narrowed down. The second unbundling can exploit
3 ERIA (2012b) proposes the concept of geographical inclusiveness, industrial inclusiveness, and societal inclusiveness. The first two are closely related to development issues and correspond to the narrowing of geographical and industrial development gaps.
26
differences in location advantages in a subtler and more articulate way than the first
unbundling.
The second is through the movement of labour at the time an industrial
agglomeration is forming. This particularly works when labour moves domestically from a
rural area to an urban area. Less-developed countries typically have a huge
agricultural/rural/informal sector where massive redundant labour resides. Smooth
labour movements from the agricultural/rural/informal sector to the non-
agricultural/urban/formal sector are often effective in reducing population below the
poverty line and at the same time providing inexpensive labour to the manufacturing and
modern services sectors.
Figure 2.12 illustrates the situation by using a simple diagram a la Lewis (1954).
OxOz stands for the total labour supply of this country, and VMPLx0 and VMPLz
0 are curves
that represent the original values of marginal product of labour (VMPL) in the rural sector
(x) and the urban sector (z).4 In the initial situation, OxL0 and OzL0 are the amount of
labour employed by sector x and sector z with equalised wages at w0. The area below the
VMPL curve corresponds to the total value of production in each sector. BA is a flat or
nearly flat portion of VMPLx0 curve that indicates redundant labour in the rural area5.
Suppose that new investment or productivity growth occurs in sector z and the VMPLz
curve shifts up to VAPLz1. If labour can move without friction, the BA portion of labour
moves from rural to urban. By this labour movement, the capitalist in sector z gains area
BCA while shifted labour earns area BAL1L0 in sector z. Here, the wage level still stays
around w0. However, if sector z has further investment or productivity growth, labour will
shift more and the upward-sloping portion of VMPLx0 will allow the wages in both sectors
to increase. This is a typical trickle-down effect from urban to rural.
The key setting here is that labour can move in a frictionless manner. In cases
where labour can move only with substantial friction, the living cost in the urban area is
substantially higher than in the rural area, education gaps are too large between rural and
urban, or the minimum wage applied in the urban area is too high, the labour movement
from rural to urban becomes smaller than BA. In the extreme, if labour cannot move at
4 The interpretation of sector x and sector z could be ‘agriculture and manufacturing’ or ‘informal and formal sectors’. 5 What redundant labour is doing or how high the marginal product of labour was a point of big debate in the literature of the 1960s, but we do not step into such an argument here.
27
all, the equilibrium for sectors x and z is A and C where capitalist in sector z loses BCA, and
even further upward shifts in VMPLz do not provide wage increases in sector x.
Figure 2.12. Labour Movements from the Informal to the Formal Sector
Source: ERIA CADP research team.
As shown in Chapter 4, ASEAN and East Asia have achieved relatively smooth
labour movements from rural to urban, from agriculture to manufacturing/services, and
from informal to formal sectors, which allow workers’ wages to stay relatively low
compared to GDP per capita. This is because economic growth has mostly been led by the
manufacturing sector and related services and educational gaps between rural and urban
have been relatively small. This has rapidly reduced the population living below the
poverty line.
Let us turn to industrial development gaps. In industrial agglomeration, plants or
establishments held by multinationals and local firms are located side by side. This means
that local firms are sitting just next to higher technology and managerial ability. This form
of globalisation certainly enhances competition, which may hurt local firms. At the same
time, it may generate opportunities for local firms to acquire new technology and
VMPLx VMPLz
Ox Oz
w0 w0
w1
L0L1
A
B
C
Investmentor
prod. growth
VMPLx0
VMPLz0
VMPLz1
28
managerial ability. If it works as intended in our development strategy, we can narrow
industrial development gaps.
2-5. Infrastructure for Connectivity and Innovation
2-5-1. The 2x3 matrix for infrastructure development
Based on our conceptual framework, infrastructure development can be tabulated
as Table 2.3. The first row refers to infrastructure for connectivity while the second
denotes infrastructure for innovation. Each of them is further classified by the degree of
involvement in production networks, i.e., Tier 1, Tier 2, and Tier 3. Since infrastructure for
Tiers 1a and 1b is often inseparable, the following will work with Tier 1 in total.
Table 2.3. Infrastructure for Connectivity and Innovation
Note: LCC = Low-cost carrier, LRT = Light rail transit. Source: ERIA CADP research team.
Tier 1:Forming industrial
agglomeration
Tier 2:Coming into production
networks
Tier 3:Rural development for
creating business
Infrastructure for connectivity
Turnpike connectivity with other industrial agglomerations- Full-scale port with
container yard/airport for regular carriers and LCC
- Multi-modal (cargo, passenger)
- Institutional connectivity for reducing transaction costs
High-grade connectivity to participate in production networks- Dual-modal (cargo,
passenger)- Capital city, border area,
connectivity grid- Mitigate border effects- Institutional
connectivity / soft infrastructure for trade facilitation
Medium-grade connectivity for various economic activities- Agriculture/food
processing, mining, labor-intensive industries, tourism, and others
Infrastructure for innovation
Metropolitan development for full-scale industrial agglomeration and urban amenities- Highway system, urban
transport (LRT, subway, airport access)
- Mass economic infrastructure services (industrial estates, electricity, energy, water, and others)
- Urban amenities to nurture/attract intellectual people
Urban/suburbandevelopment for medium-scale industrial agglomeration- Urban/suburban
development plan for a critical mass of industrial agglomeration
- Economic infrastructure services (special economic zones, electricity, water, and others)
Discovery and development of historical/cultural/ natural heritage- Premium tourism- Cultural studies
29
2-5-2. Infrastructure for connectivity
2-5-2-1. Tier 1
A full-sized industrial agglomeration requires ‘turnpike’ connectivity with other
industrial agglomerations by overcoming time and space. Expensive but essential
infrastructure includes a full-scale port with an ample container yard for main shipping
routes and a large airport for both regular carriers and low-cost carriers. Turnpike
connectivity must be multi-modal, ‘fast and slow’, and ‘high-priced and low-priced’, for
both cargoes and passengers as far as the physical geography allows.
Institutional connectivity should be achieved at a high level in order to support
efficient industrial agglomerations and affluent urban amenities. Institutional
harmonisation or convergence must be pursued to reduce transaction costs.
2-5-2-2. Tier 2
Countries/regions that are coming into production networks must establish ‘high-
grade’ connectivity. Dual-modal connectivity, i.e. fast and slow, must be provided for both
cargoes and passengers with road, port, and air transportation. Plans for middle-distance
high-speed railways should be reviewed from a viewpoint of economic viability; due to
competition with air transportation, 800–1,000 km seem to be a threshold.
The balance between the capital city and border areas must be carefully
maintained. Connectivity grids may be a key to extend connectivity to Tier 3 regions.
Connectivity with information and communications technology (ICT) would work
as both supplement and substitute for other types of connectivity. The use of ICT should
be aggressively explored.
In Tier 2, border effects are still likely to be barriers to production networks, which
should immediately be mitigated. In particular, soft infrastructure for trade facilitation is
important.
2-5-2-3. Tier 3
‘Medium-grade’ connectivity is needed for various economic activities such as
agriculture/food processing, mining, labour-intensive industries, tourism, and others. A
bit slow but reliable logistics links help various industries to be activated.
30
2-5-3. Infrastructure for innovation
2-5-3-1. Tier 1
Metropolitan development must include the construction of full-scale industrial
agglomeration in the size of 100 km diameter and urban amenities. At a higher
development stage, urban amenities are going to increase their importance. We need to
control positive and negative agglomeration effects by taking care of externalities.
The efficient highway system and urban transport, such as light rail transit,
subways, and airport access, are needed. These projects may be justified even if the
financial returns to the projects are expected to be small because they may generate huge
positive externalities and mitigate negative externalities such as traffic congestion.
Mass economic infrastructure services should also be provided; these include
industrial estates, electricity, energy, water, and others.
After reaching the upper middle–income level, urban amenities must be
emphasised in infrastructure development. Urban amenities include (i) the presence of a
rich variety of services and consumer goods, (ii) aesthetics and physical setting, (iii) good
public services, and (iv) speed (Glaeser, Kolko, and Saiz, 2001). Infrastructure is certainly
needed to achieve these.
2-5-3-2. Tier 2
Although the scale would be medium-size, i.e. 50 km diameter or so, the formation
of industrial agglomerations should be initiated. Urban/suburban development plans for
infrastructure development must be prepared in order to reach a critical mass of
economic activities. Bottlenecks in economic infrastructure services, such as SEZs,
electricity, water, and others, have to be removed.
2-5-3-3. Tier 3
In some specific places, there is potential for discovering and developing historical,
cultural, or natural heritage. In such a place, we can think of premium tourism and the
establishment of a cultural study centre.